Evaluating the biodiversity changes (BIOSAFE)¶
Overview¶
This notebook facilitates using the RiverScape Python package to evaluate the effects of RiverScape measures on biodivesity using the BIOSAFE model. If this is your first time running the notebook, we advise you to read its reference paper (Straatsma et al., 2017, including its supplementary material).
# to display all output from a cell
from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"
# see: https://ipython.readthedocs.io/en/stable/config/options/terminal.html#configtrait-InteractiveShell.ast_node_interactivity
Requirement (Python modules/packages)¶
To run this notebook, please import the following Python modules.
# import standard modules
import os
import shutil
import sys
import string
import subprocess
import time
import math
import pprint as pp
# import required modules/packages (which may require installation)
import numpy as np
import pandas as pd
import geopandas as gpd
import seaborn as sns
import matplotlib.pyplot as plt
import pcraster as pcr
Please also make sure that this notebook file is in the same folder as the RiverScape Python module files that must be loaded. You can then import the RiverScape Python module by running the following cell.
from riverscape import *
import geoviews
geoviews.extension('bokeh')
Input folder¶
The input folder is given in the following location.
# set the current directory as the root directory
try:
if root_dir == None: root_dir = os.getcwd()
except:
root_dir = os.getcwd()
# by default, the input folder is given in the following location that is relative to the root directory (the current directory)
github_input_dir_source = "../input_files/input/bio/"
input_dir_source = os.path.join(root_dir, github_input_dir_source)
You may also want to set the input folder, interactively, by running the following cells.
# go to the above suggested 'input_dir_source'
os.chdir(input_dir_source); input_dir_source = os.getcwd()
# open an interactive window
get_input_dir_source = ""
# get_input_dir_source = select_directory()
# - if cancel or not used, we use the above suggested 'input_dir_source'
if get_input_dir_source == "" or get_input_dir_source == "()": get_input_dir_source = input_dir_source
input_dir_source = get_input_dir_source
msg = "The input folder is set on : {}".format(input_dir_source)
print(msg)
# return to the root folder for further processes.
os.chdir(root_dir)
Output folder¶
Please set your output folder. Please adjust if necessary.
output_folder = "/scratch-shared/edwinhs/riverscape/biosafe_development/test_biosafe_5/"
You can also set the output folder, interactively, by running the following cells.
# open an interactive window
get_output_folder = ""
# get_output_folder = select_directory()
# - if cancel or not used, we use the above defined 'output_folder'
if get_output_folder == () or get_output_folder == "": get_output_folder = output_folder
output_folder = get_output_folder
msg = "The output folder is set on : {}".format(output_folder)
print(msg)
The files in the input folder will be copied to the output folder. The output folder will also be set as the working directory. All plots would be saved in this directory.
# safe copying method
source = os.listdir(input_dir_source)
destination = output_folder
for files in source:
shutil.copy(os.path.join(input_dir_source, files), destination)
#~ # dangerous copying method
#~ if os.path.exists(output_folder): shutil.rmtree(output_folder)
#~ shutil.copytree(input_dir_source, input_dir_source)
# use the output folder as the working directory
os.chdir(output_folder)
Reading input files¶
To calculate biodiversity indices of protected and endangered species that are characteristic of the fluvial environment, this notebook applies the Spreadsheet Application For Evaluation of BIOdiversity (BIOSAFE) model (see e.g. de Nooij et al., 2003). The original spreadsheet file in an excel format is provided in the following location. We adopt the excel file to this notebook.
# the original excel file
excelFile = os.path.join(output_folder, 'BIOSAFE_20150629.xlsx')
msg = "The original excel file is given on {}".format(excelFile)
print(msg)
We adopt the spreadsheet/excel file to this notebook.
The first input to the spreadsheet is a matrix legalWeights (Legislative Weights) of user-specified weights for 16 different policy and legislative documents. The 16 legal documents considered are the most recent versions of: Dutch Red List species that are based on the criteria of the International Union for Conservation of Nature (IUCN), Dutch Flora and Fauna Act, EU Habitat Directive, EU Birds Directive, Bern Convention, and Bonn Convention. We assumed equal weights, attributing equal value to the protection by each of the legal and policy documents.
Please execute the following cells to load and visualize the legalWeights used in this study.
legalWeights = pd.read_csv(os.path.join(output_folder, 'legalWeights.csv'), index_col = 0)
bsIO.show_full_data_frame(legalWeights)
The second input is linksLaw (Legal Links), which is a relational matrix that links species(s) along the rows to the relevant policy and legislative documents (legalWeights) along the columns. If a species is protected by or mentioned in a specific law, the corresponding cell in the matrix is given a value of one, else the cell is zero.
Please execute the following cells to load and visualize the linksLaw used in this study.
linksLaw = pd.read_csv(os.path.join(output_folder, 'linksLaw.csv'), index_col = 0)
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(linksLaw)
The third input is linksEco (Ecotope Links), which is the relational matrix between species and ecotopes. The matrix is based on expert ecological judgement of species and their hydromorphological and vegetation habitat needs in different life stages.
Please execute the following cells to load and visualize the linksEco used in this study.
linksEco = pd.read_csv(os.path.join(output_folder, 'linksEco.csv'), index_col = 0)
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(linksEco)
The column titles in linksEco represent the ecotope classes used in this study. In addition to them, we introduce their merged/main classes based on the following relationship matrix lut that can be loaded and visualized by executing the following notebook cells.
#~ # using excel file
#~ excelFile = os.path.join(output_folder, 'BIOSAFE_20150629.xlsx')
#~ lut = pd.read_excel(excelFile, sheet_name = 'lut_RWES').fillna(method='ffill')
# using csv file
csvFile = os.path.join(input_dir_source, 'lut_RWES.csv')
lut = pd.read_csv(csvFile, index_col = 0)
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(lut)
In the matrix lut, the column oldEcotope represents the ecotope classes given along the top rows of linksEco, while the classes in the column newEcotope are their merged ecotope classes. Based on lut (relation between oldEcotope and newEcotope) and by running the following notebook cells, we can derive and visualize the Ecotope Links linksNewEcoAgg, consisting oldEcotope and newEcotope.
linksNewEcoAgg = biosafe.aggregateEcotopes(linksEco, lut)
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(linksNewEcoAgg)
print(linksEco.shape)
print(linksNewEcoAgg.shape)
Understanding/testing BIOSAFE in a simple instance (non-spatial mode)¶
To understand the BIOSAFE model, we introduce the following exercise running the BIOSAFE model in a simple single instance, without relevant spatial information.
For this exercise, we randomly draw the presence of each species by running the following cells. The assumption drawn is given in the table, particularly in the column speciesPresence, with the values 0 and 1 indicating absent and present.
speciesPresence = pd.DataFrame(np.random.randint(2, size=len(linksLaw)),\
columns=['speciesPresence'], \
index=linksLaw.index)
speciesPresenceFull = speciesPresence
# TODO: Menno has to define a method to simplify, to reduce the number of specieses.
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(speciesPresence)
For this exercise, we simplify all areas of ecotope classes (ecotopeArea) equal to 1e5 m2. This is done by running the following cell.
ecotopeArea = pd.DataFrame(np.ones(len(linksNewEcoAgg.columns)-1) * 1e5,\
columns = ['area_m2'],\
index = linksNewEcoAgg.columns.values[0:-1])
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(ecotopeArea)
Running the BIOSAFE in a simple single instance (non-spatial mode)¶
Given the inputs legalWeights, linksLaw and linksNewEcoAgg, as well as the assumptions speciesPresence and ecotopeArea defined above, the BIOSAFE model is initiated by running the following cell.
bs = biosafe.biosafe(legalWeights, linksLaw, linksNewEcoAgg, speciesPresence, ecotopeArea)
Then, we can calculate all BIOSAFE scores (potential and actual ones).
SScoresSpecies = bs.SScoresSpecies()
summarySScores = bs.taxGroupSums()
SEcoPot = bs.SEcoPot()
SEcoAct = bs.SEcoAct()
PTE = bs.PTE()
ATE = bs.ATE()
TES = bs.TES()
TEI = bs.TEI()
ATEI = bs.ATEI()
TFI = bs.TFI()
FI = bs.FI()
ATFI = bs.ATFI()
AFI = bs.AFI()
FIS = bs.FIS()
TFIS = bs.TFIS()
TFHS = bs.TFHS()
FTEI = bs.FTEI()
Exploring the BIOSAFE scores/indices¶
After running the BIOSAFE model, we would explore the BIOSAFE scores/indices by running the following cells. Note that we distinguish between potential biodiversity indices, which depend on ecotope presence only, and actual biodiversity indices, which combine ecotope and species presences. To understand the description of each score, we refer to the supplementary material of Straatsma et al. (2017) that provides all equations to calculate these scores. Figure S1 in the supplementary material shows the flow chart of the BIOSAFE methodology, while a graphical illustration of BIOSAFE score computation is given in Figure S2.
Note that during this exploration, you may want to skip some cells for visualizing some values in data frames or clear their outputs after executing them, particularly if they involve large data frames.
We start this exploration with SScoresSpecies that consists of the calculated potential and actual species scores, Spotential and Sactual (PSS and ASS in the paper). For each species, their calculated values are given as follows.
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(SScoresSpecies)
Based on SPotential and SActual, the potential and actual taxonomic group scores, PTB and ATB, can be calculated for each taxonomic group as follows. In addition, the taxonomic group biodiversity, TBS, saturation can also be calculated.
bsIO.show_full_data_frame(summarySScores)
For each ecotope class, we can also compute the potential and actual species specific scores for each ecotope. Their values for each species and ecotope areas are given in the following two cells, the first one shows their potential scores, SEcoPot, while the latter contains their actual scores, SEcoAct.
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(SEcoPot)
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
# bsIO.show_full_data_frame(SEcoAct)
For each taxonomic group and for each ecotope class, the sum of SEcoPot can be calculated as PTE, potential taxonomic group ecotope constant.
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
bsIO.show_full_data_frame(PTE)
For each taxonomic group and for each ecotope class, the sum of SEcoAct can be calculated as ATE, actual taxonomic group ecotope constant.
# Note: This cell is for visualizing data only and may be heavy.
# You may want to skip it or clear its output after executing it.
bsIO.show_full_data_frame(ATE)
Based on ATE and PTE, we can calculate TES, taxonomic group ecotope saturation index, for each taxonomic group and ecotope area.
bsIO.show_full_data_frame(TES)
Based on PTB and PTE, we can calculate TEI, taxonomic group ecotope importance constant, for each taxonomic group and ecotope area.
bsIO.show_full_data_frame(TEI)
Based on TEI and TES, we can calculate ATEI (ATE in the paper), actual taxonomic group ecotope constant, for each taxonomic group and ecotope area.
bsIO.show_full_data_frame(ATEI)
Based on the ecotopeArea assumed above and TEI, the potential taxonomic group flood plain importance score, TFI (PotTax in the paper), can be calculated as follows.
bsIO.show_full_data_frame(TFI)
The sum of TFI is reported as FI (PotAll in the paper), the potential floodplain importance score.
bsIO.show_full_data_frame(FI)
Based on the ecotopeArea assumed above and ATEI, the actual taxonomic group flood plain importance score, ATFI (ActTax in the paper), can be calculated as follows.
bsIO.show_full_data_frame(ATFI)
The sum of ATFI is reported as the actual floodplain importance score, AFI (ActAll in the paper).
bsIO.show_full_data_frame(AFI)
Based on AFI and FI, we can calculated the floodplain importance saturation, FIS (SatAll in the paper), as follows.
bsIO.show_full_data_frame(FIS)
For each taxonomic group, the taxonomic group floodplain importance saturation, TFIS (SatTax in the paper), can be calculated based on ATFI and FIS. TFIS describes the fraction/saturation of actual over potential biodiversity.
bsIO.show_full_data_frame(TFIS)
Based on TEI and ecotopeArea, we can calculate TFHS (HabDiv in the paper), the taxonomic group floodplain habitat saturation, which describes the fraction of suitable ecotopes, weighed by law, for each taxonomic group.
bsIO.show_full_data_frame(TFHS)
# bsIO.show_full_data_frame(FTEI)
# FTEI is not found in the paper.
# FTEI: Floodplain Taxonomic group Ecotope Importance
Running the BIOSAFE in a spatial mode¶
In the following cells, we will demonstrate the BIOSAFE model in a spatial model. The BIOSAFE model will be fed by spatial data, particularly on ecotope distribution and species monitoring data.
We assume the following floodplain section map flpl_sections, which would be loaded and vizualized by executing the following notebook cells.
flpl_sections_f = os.path.join(output_folder, 'flpl_sections.map')
flpl_sections = pcr.readmap(flpl_sections_f)
plot(flpl_sections)